Key takeaway
- How to automate Drupal releases on Azure with continuous delivery: pipeline stages, where AI agents help, and what to check before you adopt it.
Continuous delivery for Drupal on Azure is a release practice in which every approved code change is built, tested and deployed through an automated Azure DevOps pipeline, rather than by hand on a server. Drupal is open source content management software, and a pipeline gives its releases a repeatable path from merge request to production.
This page is for engineering leads and platform teams who run Drupal for an enterprise or public sector site. It covers the problem, the pipeline steps, where AI agents fit, a before and after comparison, and what to check before adopting it.
What is continuous delivery for a Drupal site?
The Azure Pipelines documentation defines continuous delivery as "the process of building, testing, and deploying code to one or more test or production environments". It builds on continuous integration, where code changes are merged often and checked automatically. For Drupal, that means the custom modules, themes and configuration of a site move through the same automated route every time.
Drupal describes itself as content management software and says its flexibility comes from modularity. That modularity is also why a release can be complicated: code, configuration and database updates all have to land together.
What problem does a Drupal release pipeline solve?
Teams that moved Drupal from traditional physical servers to the cloud often keep the old release habits: someone logs in, pulls code, runs updates and restarts the web server. The problem is not the tools. It is that the steps live in people's heads, so releases are slow, hard to repeat and hard to audit.
A pipeline replaces those habits with a defined flow. The goal is to automate the software release process through to the production environment on Microsoft Azure, with an approval step where a person decides.
How does the delivery pipeline work, step by step?
A typical flow for a production release looks like this. Adjust each step to your own hosting and branching model.
- A developer opens a merge request to the release branch.
- The pipeline builds the code and runs automated checks: Drupal coding standards, unit tests and kernel tests.
- An administrator reviews and approves the merge request.
- The release branch is checked out and the latest code is pulled into the target workspace.
- The database is updated so it matches the new code.
- The web service, such as Apache, is restarted.
- The new code is live in production, and reports from the run are stored for review.
The Azure Pipelines documentation says targets include "virtual machines, environments, containers, on-premises and cloud platforms". Choose the target that matches how you host Drupal today, then keep the pipeline definition in version control beside the code.
Why pair DevOps with microservices for Drupal?
Many teams move supporting functions out of the Drupal codebase into separate services, such as search, integrations or media handling. Each service can then be scaled, deployed and owned on its own. If one fails, the fault stays inside it. A domain-driven design helps decide where the boundaries sit, and each service keeps its own data so privacy rules are easier to apply.
Where do AI agents fit in Drupal delivery?
Agents suit the work around the pipeline, not the approval itself. They can read failed check output and summarise the likely cause, draft release notes from merged changes, and flag a missing database update before a person approves. ElixirClaw is XenonStack's AI agents computing platform. Whichever platform you use, give an agent only the access it needs, and keep production approval with people.
How does manual release compare with an automated pipeline?
| Area | Manual release | Automated pipeline |
|---|---|---|
| Steps | Run by hand on the server | Defined once in version control and repeated |
| Checks | Often skipped under time pressure | Coding standards and tests run on every merge request |
| Approval | Informal, by message | Recorded approval before production |
| Database updates | Easy to forget | A fixed step in the flow |
| Audit trail | Scattered notes | Run history and reports in one place |
What should you check before adopting it?
- Is Drupal code, theme and configuration all in version control?
- Do you have automated tests, or only manual checks, today?
- Can you roll back a release, including its database changes?
- Who approves production releases, and is that recorded?
- Where do secrets such as database credentials live?
- Which environments exist between a developer machine and production?
To map one decision your agents make, and the data and policy it depends on, book a Context Workshop. See what happens in the workshop.
Frequently asked questions
What is continuous delivery for Drupal?
It is a practice where every approved change to Drupal code is built, tested and prepared for release through an automated pipeline, so releases follow the same repeatable steps each time instead of manual server work.
Does Azure Pipelines support Drupal?
Azure Pipelines builds, tests and deploys code from version control, including PHP projects, to targets such as virtual machines, containers, and cloud platforms. A Drupal project is PHP code, so it can run through a pipeline that you define for its build, test and deploy steps.
Why use microservices with Drupal?
Splitting supporting functions into separate services lets teams scale, deploy and isolate faults for each service on its own. Drupal itself can remain the content platform while those services evolve at their own pace.
Where do AI agents help in a delivery pipeline?
Agents can read pipeline results, summarise failed checks, and prepare release notes for a person to review. Approval for production stays with people, and every agent action should be logged.
What should we set up first?
Start with one site and one pipeline: build, code standard checks and tests on every merge request, then a gated deployment to a test environment. Add production after that path is stable.
Ready to put AI agents to work? See an ElixirClaw demo: the agentic OS to build, orchestrate and govern AI agents across your enterprise systems.
Explore the XenonStack platform family: ElixirData · ElixirClaw · ElixirHub · NexaStack · MetaSecure · Neural AI · AgentGRC · AgentSRE · AgentAnalyst · AgentInstruct · AgentRAI · AgentScrum · AgentSearch · AgentForce · AI Aviator · SAIF Aviator · CohortIQ · XenonStack AI · XenonStack UAE